Pre-cluster symptoms in a Taiwanese cohort of cluster headache: symptom profiles and clinical predictions

Jr‐Wei Wu, Shu‐Ting Chen, Yen-Feng Wang, Shih‐Pin Chen, S.-h. Tseng, Yih-Shiuan Kuo, Wei‐Ta Chen, Chia‐Chun Chiang, Shuu‐Jiun WangThe Journal of Headache and Pain 2024 · DOI

Abstract

Pre-cluster symptoms (PCSs) are highly prevalent in patients with cluster headache (CH) and serve as significant clinical markers. This study demonstrates that PCSs can accurately predict the onset of upcoming headache bouts and identify patients likely to respond to first-line preventative treatments like verapamil. Different symptom categories of PCSs appear to have distinct clinical implications for disease management.

Introduction

Cluster headache (CH) is one of the most severe pain conditions known, characterized by cyclical "bouts" of intense unilateral pain. A significant challenge in clinical practice is the reactive nature of treatment; patients typically seek help only after a bout has commenced. Identifying a "therapeutic window"—a period of warning signs before the pain begins—could allow for pre-emptive intervention. This study aimed to profile these pre-cluster symptoms (PCSs) and determine if they could predict both the onset of a bout and the efficacy of verapamil, the standard preventative therapy.

Methods

Researchers conducted a prospective study at the Taipei Veterans General Hospital involving 168 patients diagnosed with CH according to ICHD-3 criteria. The cohort was predominantly male (76.8%) with a high prevalence of episodic CH (97.6%).

The study was structured into two primary phases: a symptom profiling sub-study and a treatment prediction sub-study. During active bouts, patients provided detailed accounts of symptoms experienced in the week prior to the bout's onset, covering 34 specific symptoms across seven categories. To assess treatment response, patients were started on verapamil alongside a 14-day steroid bridge (transitional therapy). Clinical outcomes were tracked via headache diaries, with "responders" defined as those achieving a $\ge$50% reduction in weekly attack frequency after four weeks.

Overview of the study design, including screening, intervention with verapamil and transitional therapy, and follow-up phases.

Results

Prevalence and Predictability

PCSs were remarkably common, reported by 88.7% of the cohort. The median lead time for these symptoms was 24 hours before the bout began, though the range extended up to 20 days. Notably, 57.7% of patients with PCSs could accurately predict an upcoming bout based on these signals.

Symptom Profiles as Red Flags

Head and neck pain was the most frequent PCS (81%) and the strongest predictor of an imminent bout (Odds Ratio [OR] = 4.0). Specific sensations, such as a "dull" feeling in the usual headache area, were highly reliable indicators. Other common symptoms included constitutional issues (77.4%) and mood or fatigue changes (69%). Restlessness was also identified as a significant predictor (OR = 3.3).

Predicting Treatment Response

The study found that specific PCS phenotypes correlate with better outcomes on verapamil. Patients experiencing sleep alterations (OR = 2.5) or at least one cranial autonomic symptom (such as lacrimation or nasal congestion) during the pre-cluster phase (OR = 2.7) were significantly more likely to be responders to verapamil. Furthermore, patients with predictable bouts often exhibited higher circadian rhythmicity in their attacks, suggesting a strong link between PCSs and hypothalamic involvement.

Discussion

These findings suggest that CH is not a monolithic condition and that the pre-cluster phase offers vital diagnostic data. The ability to predict a bout through head/neck pain or restlessness provides a window for early medical intervention. Moreover, using sleep and autonomic symptoms as biomarkers for verapamil response could help clinicians personalize treatment plans more effectively.

Limitations of the study include potential recall bias due to the retrospective nature of symptom reporting during bouts and the specific focus on a Taiwanese population, which may limit generalizability to other ethnic groups. Future research should investigate whether initiating treatment immediately upon the appearance of PCSs can successfully abort or shorten a cluster bout.

Summary updated

Abstract

PCSs were very common in CH and could be used to predict upcoming bouts. Different symptom categories of PCSs may have different clinical implications.
BibTeX
@article{wu2024pre,
  author = {Jr‐Wei Wu and Shu‐Ting Chen and Yen-Feng Wang and Shih‐Pin Chen and S.-H. Tseng and Yih-Shiuan Kuo and Wei‐Ta Chen and Chia‐Chun Chiang and Shuu‐Jiun Wang},
  title = {{Pre-cluster symptoms in a Taiwanese cohort of cluster headache: symptom profiles and clinical predictions}},
  journal = {The Journal of Headache and Pain},
  year = {2024},
  doi = {10.1186/s10194-024-01862-8},
}
Figure